Chemical Interface Damping (CID)

• Chemical Interface Damping (CID) is the enhanced electron collision frequency at chemically modified metal interfaces, leading to increased plasmon damping and broadened resonances.
• CID is distinguished from geometric surface scattering because it arises from chemical-induced interfacial roughness and direct charge-transfer pathways.
• Experimental studies using spectroscopic ellipsometry and waveguide methods quantify CID’s spectral, resistive, and hot-carrier generation impacts in plasmonic systems.

Searching arXiv for recent CID papers and related metal–molecule damping work. Chemical interface damping (CID) is the increase of the effective collision frequency $\gamma$ of conduction electrons in a metal that occurs when the metal’s interface is chemically modified. In plasmonic and intraband optical regimes, this increase in $\gamma$ raises the dissipative part of the metal response, broadens plasmon resonances, and alters optical absorption, propagation loss, and hot-carrier pathways. Across recent work, CID is treated as an interfacial damping channel distinct from purely geometric size-dependent surface scattering: chemistry can create an effective roughness that enhances momentum randomization, and it can open direct charge-transfer pathways into hybrid interfacial states or adsorbate states (Pfeiffer et al., 5 Sep 2025). Experiments on thiol self-assembled monolayers, plasmonic waveguides with molecular adsorbates, and electrochemically oxidized Au have established that CID can be spectrally structured, molecule-specific, and quantitatively linked to both optical linewidths and DC transport (Stefancu et al., 16 Jul 2025, Pfeiffer et al., 2024).

1. Conceptual definition and placement among damping channels

In plasmonic nanostructures, a surface plasmon resonance decays through several channels: radiative damping, bulk nonradiative losses, surface scattering, and chemical interface damping. Within this decomposition, CID is an additional nonradiative decay channel activated only when molecules are adsorbed or when the metal interface is otherwise chemically altered; the plasmon then couples directly to electronic degrees of freedom at the metal–molecule or metal–oxide interface and loses energy through interfacial electron transfer or scattering (Stefancu et al., 16 Jul 2025).

A compact statement of this partition is

$$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$

with plasmon quality factor

$Q=\omega/(2\Gamma_{\mathrm{tot}}).$

In waveguide language, the corresponding propagation loss coefficient is written as

$$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$

where $\alpha_i$ is the intrinsic loss of the bare waveguide (Stefancu et al., 16 Jul 2025).

CID is distinguished from size-dependent surface scattering in two ways. First, size-dependent scattering is purely geometric: when characteristic dimensions fall below the bulk mean free path, electrons collide more frequently with a physical boundary and $\gamma$ increases roughly with $1/L$. By contrast, CID changes the nature of the surface collision itself through chemistry. Second, CID need not involve any change in physical dimensions; instead, the interface becomes a more efficient sink for electron momentum because of interfacial roughness, hybridization, or charge-transfer pathways (Pfeiffer et al., 2024).

This distinction is especially important for planar Au systems. Electrochemical oxidation of Au and adsorption of alkanethiol self-assembled monolayers both increase $\gamma$ without changing the macroscopic geometry of the film, which indicates that interfacial chemistry alone can measurably modify plasmonic damping (Pfeiffer et al., 5 Sep 2025, Pfeiffer et al., 2024).

2. Electrodynamic description and optical observables

Within the Drude picture, intraband optical absorption arises when conduction electrons undergo momentum-relaxing collisions with phonons, defects, grain boundaries, and surfaces; CID adds new interfacial scattering channels and interfacial charge-transfer pathways. The metal dielectric function can be written as

$$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$

and the associated intraband conductivity is

$\gamma$0

An increase in $\gamma$1 therefore directly increases $\gamma$2 and $\gamma$3 at optical frequencies, broadening and damping plasmon resonances (Pfeiffer et al., 5 Sep 2025).

For Au in the near-IR and visible, the free-electron response is commonly embedded in a Drude–Lorentz form to include the onset of interband transitions:

$\gamma$4

with a Lorentz oscillator

$\gamma$5

describing the $\gamma$6-band to $\gamma$7-band transitions that begin near $\gamma$8 (Pfeiffer et al., 5 Sep 2025).

Spectroscopic ellipsometry provides a direct route to extracting $\gamma$9. It measures the complex reflectance ratio

$$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$0

where $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$1 and $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$2 encode the amplitude ratio and phase difference of $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$3- and $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$4-polarized reflected light. For a single interface,

$$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$5

with

$$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$6

For layered stacks, the reflection coefficients are computed with a standard transfer-matrix formalism (Pfeiffer et al., 5 Sep 2025).

Electrochemical ellipsometry on Au oxidation exploits a useful sensitivity separation: $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$7 is most sensitive to $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$8, while $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$9 is most sensitive to oxide thickness $Q=\omega/(2\Gamma_{\mathrm{tot}}).$0. In the oxidation study, the Au plasma frequency $Q=\omega/(2\Gamma_{\mathrm{tot}}).$1 and high-frequency permittivity $Q=\omega/(2\Gamma_{\mathrm{tot}}).$2 were fitted once at low potential and then held fixed during cycling, while only $Q=\omega/(2\Gamma_{\mathrm{tot}}).$3 and the oxide effective thickness were varied (Pfeiffer et al., 2024).

In waveguide measurements, CID was quantified from transmission loss by a cut-back relation,

$Q=\omega/(2\Gamma_{\mathrm{tot}}).$4

followed by

$Q=\omega/(2\Gamma_{\mathrm{tot}}).$5

with $Q=\omega/(2\Gamma_{\mathrm{tot}}).$6 the SPP group velocity (Stefancu et al., 16 Jul 2025).

3. Microscopic mechanisms and CID regimes

Recent work identifies two distinct CID regimes. One is direct electronic transition into molecular acceptor states, exemplified by biphenyl thiol (BPT). The other is nonresonant inelastic interfacial scattering, exemplified by ATP, adenine, DDT, and the low-energy component of decanethiol-induced CID on Au(111) (Stefancu et al., 16 Jul 2025, Pfeiffer et al., 5 Sep 2025).

For the resonant regime, the mechanism is coherent, one-step charge transfer driven by the plasmon near field from occupied metal states around the Fermi level $Q=\omega/(2\Gamma_{\mathrm{tot}}).$7 into hybridized metal–molecule acceptor states, identified in the waveguide study with the molecule’s LUMO. The resonance condition is

$Q=\omega/(2\Gamma_{\mathrm{tot}}).$8

BPT has its LUMO centered at approximately $Q=\omega/(2\Gamma_{\mathrm{tot}}).$9 above $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$0, within reach of the SPR energies used, and its CID shows strong plasmon-energy dependence (Stefancu et al., 16 Jul 2025). A rate picture based on Fermi’s Golden Rule is written as

$$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$1

with this interfacial transfer channel contributing to $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$2 (Stefancu et al., 16 Jul 2025).

For the nonresonant regime, plasmon damping proceeds via interfacial inelastic scattering of conduction electrons into adsorbate degrees of freedom, including nonadiabatic electron–vibration coupling and diffuse scattering mediated by transient coupling to unoccupied molecular states, without requiring a discrete resonant transition between hybridized states (Stefancu et al., 16 Jul 2025). The microscopic picture is that adsorbates disrupt translational symmetry, so electrons near $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$3 undergo diffuse, inelastic scattering at the interface. Because those same near-$$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$4 electrons also carry DC current, the same mechanism contributes to adsorbate-induced changes in DC resistivity (Stefancu et al., 16 Jul 2025).

The decanethiol/Au(111) ellipsometry study resolves a related two-component structure in energy space. At low photon energies, the CID increment is approximately constant and attributed to induced roughness and added scattering centers at the interface. Above a threshold near $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$5, the CID increment increases approximately linearly with photon energy and is attributed to direct charge transfer from occupied hybrid Au–S states into Au $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$6 states near $$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$7 (Pfeiffer et al., 5 Sep 2025). A compact phenomenological description is

$$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$8

with

$$\alpha=\alpha_{\mathrm{bulk}}+\alpha_{\mathrm{rad}}+\alpha_{\mathrm{surf}}+\alpha_{\mathrm{CID}}\equiv \alpha_i+\alpha_{\mathrm{CID}},$$9

and

$\alpha_i$0

In that system, $\alpha_i$1, $\alpha_i$2, and $\alpha_i$3 were estimated from pointwise fits, although no unique global $\alpha_i$4 was imposed in the analysis (Pfeiffer et al., 5 Sep 2025).

Mechanistic support for the $\alpha_i$5 onset comes from spin-polarized DFT on the Au–thiol headgroup. Using methanethiol as a proxy for decanethiol, projected DOS showed hybrid Au–S HOMO-derived states centered approximately $\alpha_i$6 below $\alpha_i$7 on S and on Au atoms bonded to S, while bulk-like Au showed the $\alpha_i$8-band approximately $\alpha_i$9 below $\gamma$0 (Pfeiffer et al., 5 Sep 2025). Bader analysis indicated small charge accumulation on S, approximately $\gamma$1 to $\gamma$2, consistent with covalent Au–S bonding and hybridization that generates interfacial states. Structures including Au adatoms shifted the dominant S-related states deeper, approximately $\gamma$3, and weakened the $\gamma$4 feature; these did not match the observed onset as well, supporting a bridge-like adsorption on flat Au(111) under the reported conditions (Pfeiffer et al., 5 Sep 2025).

The oxidation study leaves the roughness-versus-charge-transfer decomposition open. It explicitly notes that the spectral window of $\gamma$5–$\gamma$6 precludes identifying energy thresholds for interfacial charge transfer and that distinguishing roughness-induced CID from charge-transfer-induced CID remains an open question (Pfeiffer et al., 2024). This suggests that broad spectral access is essential if one seeks to isolate different microscopic CID channels from ellipsometric data alone.

4. Experimental realizations and quantitative behavior

Three recent experimental realizations define the present quantitative picture of CID on Au.

System	Method	Key quantitative outcome
1-decanethiol SAM on template-stripped Au(111)	Broadband spectroscopic ellipsometry with pointwise extraction of $\gamma$7	$\gamma$8; onset $\gamma$9; $1/L$0 (Pfeiffer et al., 5 Sep 2025)
ATP, BPT, DDT, adenine on Au gap-plasmon waveguides and thin films	Optical cut-back loss plus 4-point resistivity	Two regimes: BPT shows $1/L$1 CID decrease at $1/L$2 vs $1/L$3; ATP shows negligible change (Stefancu et al., 16 Jul 2025)
Electrochemical Au oxidation on Au(111) and polycrystalline Au	In-situ electrochemical ellipsometry	Maximum $1/L$4 on single-crystal Au(111); linear growth with oxide thickness up to $1/L$5, then saturation (Pfeiffer et al., 2024)

For decanethiol on Au(111), the substrate was a $1/L$6 template-stripped Au nanolayer predominantly Au(111) ($1/L$7) on glass with manufacturer STM RMS roughness $1/L$8. The SAM was formed by immersion in $1/L$9 1-decanethiol in ethanol for $\gamma$0, followed by rinsing in ethanol and isopropanol. Ellipsometric fitting yielded decanethiol layer thicknesses of $\gamma$1 and $\gamma$2 on two samples, consistent with literature values $\gamma$3–$\gamma$4; the refractive index of the SAM was taken as $\gamma$5 (Pfeiffer et al., 5 Sep 2025). Measurements used a J.A. Woollam RC2 with parallel-beam optics at five incidence angles, $\gamma$6, $\gamma$7, $\gamma$8, $\gamma$9, and $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$0, over the full instrument range $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$1–$$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$2, although modeling was restricted to $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$3–$$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$4 to avoid strong interband masking (Pfeiffer et al., 5 Sep 2025).

The modeling workflow in that study proceeded in three steps. First, a spectral fit on bare Au(111) over $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$5–$$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$6 with Drude–Lorentz dispersion included an energy-dependent electron–electron scattering term,

$$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$7

and one Lorentz oscillator for the incipient $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$8 interband transition. This yielded $$\varepsilon(\omega)=\varepsilon_\infty-\frac{\omega_p^2}{\omega(\omega+i\gamma)} =\varepsilon_\infty-\frac{\omega_p^2}{\omega^2+i\gamma\omega},$$9 and $\gamma$00, in excellent agreement with theoretical expectations $\gamma$01 and ultrafast measurements (Pfeiffer et al., 5 Sep 2025). Second, a narrowband fit in $\gamma$02–$\gamma$03 on the thiol-coated sample assumed spectrally constant CID in that band to extract SAM thickness. Third, pointwise fits were performed with $\gamma$04, $\gamma$05, $\gamma$06, and $\gamma$07 fixed from the bare-Au fit, yielding

$\gamma$08

At low photon energies $\gamma$09–$\gamma$10, the angle-averaged $\gamma$11 was approximately $\gamma$12 and essentially constant versus energy. Above the threshold near $\gamma$13, it rose approximately linearly, reaching approximately $\gamma$14 at $\gamma$15 (Pfeiffer et al., 5 Sep 2025).

The waveguide and resistivity study used strongly confined gap-plasmon waveguides fabricated from Au ($\gamma$16 Au with $\gamma$17 Cr adhesion) on borosilicate glass by e-beam lithography. Gap widths were below $\gamma$18, with approximately $\gamma$19 free-space to SPP coupling efficiency. Continuous-wave excitation was applied at $\gamma$20 ($\gamma$21), with additional measurements at $\gamma$22 ($\gamma$23) (Stefancu et al., 16 Jul 2025). Adsorption on Au films and waveguides was from $\gamma$24 solutions. Thiol monolayers formed within the first approximately $\gamma$25 due to strong Au–S bonding, while adenine required approximately $\gamma$26 to reach a resistivity plateau (Stefancu et al., 16 Jul 2025). The extracted CID rates followed the trend ATP $\gamma$27 BPT $\gamma$28 DDT $\gamma$29 adenine, and the representative ratio $\gamma$30 experimentally, versus approximately $\gamma$31 theoretically, corresponded to a discrepancy of about $\gamma$32 (Stefancu et al., 16 Jul 2025).

For electrochemical oxidation, template-stripped single-crystal Au(111) nanolayers ($\gamma$33) on glass chips and sputtered $\gamma$34 polycrystalline Au on Si with Ti adhesion were studied in degassed $\gamma$35 $\gamma$36 in a three-electrode configuration (Pfeiffer et al., 2024). Cyclic voltammetry ran between $\gamma$37 and $\gamma$38 vs RHE, extended to $\gamma$39 for thicker oxides. Au(111) showed an oxidation peak at $\gamma$40 and reduction at $\gamma$41 (Pfeiffer et al., 2024). In the intraband window $\gamma$42–$\gamma$43, single-crystal Au(111) exhibited a linear increase of $\gamma$44 with effective oxide thickness up to $\gamma$45 and then saturation, with maximum $\gamma$46 at $\gamma$47. Scans limited to $\gamma$48 gave $\gamma$49 at $\gamma$50 (Pfeiffer et al., 2024). Polycrystalline Au showed similar CID magnitudes, approximately $\gamma$51, but with baseline evolution consistent with increased roughness or residual oxygen in grains (Pfeiffer et al., 2024).

5. Relation to DC resistivity, scattering cross-sections, and normalization

A major development in CID research is the explicit linkage between optical damping and DC transport. In a Drude-like conductivity model,

$\gamma$52

so the DC resistivity is

$\gamma$53

Using Matthiessen’s rule,

$\gamma$54

In the adsorbate study, the same interfacial scattering channel that increases DC resistivity also increases optical damping at SPR frequencies, producing $\gamma$55 (Stefancu et al., 16 Jul 2025).

The DC diffuse scattering cross-section was obtained from the initial slope of resistivity versus adsorbate coverage:

$\gamma$56

where $\gamma$57 is film thickness and $\gamma$58 is the number of adsorbates per unit area (Stefancu et al., 16 Jul 2025). Quantitatively, the reported DC resistivity change per monolayer was $\gamma$59 for ATP, $\gamma$60 for BPT, $\gamma$61 for DDT, and $\gamma$62 for adenine. The corresponding DC diffuse scattering cross-sections were $\gamma$63 for ATP, $\gamma$64 for BPT, $\gamma$65 for DDT, and $\gamma$66 for adenine (Stefancu et al., 16 Jul 2025).

The same study reported that CID rates correlate strongly with the adsorbate-induced DC scattering cross-section after subtracting the perpendicular dipole contribution,

$\gamma$67

This correlation indicates a common origin in interfacial scattering for the nonresonant regime (Stefancu et al., 16 Jul 2025). The perpendicular dipole moments obtained from DFT were $\gamma$68 for BPT, $\gamma$69 for ATP, and $\gamma$70 for DDT, with adenine much smaller; the empirical dipole contribution to DC scattering was inferred as approximately $\gamma$71–$\gamma$72 per Debye (Stefancu et al., 16 Jul 2025).

A different normalization is used for comparing CID strengths across planar films and nanoparticles. The oxidation study defines

$\gamma$73

where $\gamma$74 is an effective mean free path to the surface. For convex nanoparticles, $\gamma$75. For planar Au, using penetration depth $\gamma$76 at $\gamma$77 gives $\gamma$78 (Pfeiffer et al., 2024). Using this normalization, electrochemical oxidation of single-crystal Au(111) reached $\gamma$79, exceeding reported values for thiol-functionalized Au nanorods ($\gamma$80–$\gamma$81) and for ALD oxide coatings on Au nanoparticles (TiO$\gamma$82: $\gamma$83; HfO$\gamma$84: $\gamma$85; Al$\gamma$86O$\gamma$87: $\gamma$88) (Pfeiffer et al., 2024).

This normalization matters because raw $\gamma$89 values depend on how frequently optically excited electrons encounter the interface. A plausible implication is that comparisons of CID magnitude across geometries are most meaningful when scaled by interface-encounter statistics rather than by linewidth change alone.

6. Comparative interpretation, implications, and unresolved issues

Several comparative conclusions emerge from the available studies. First, CID is not spectrally uniform in general. Decanethiol on Au(111) shows a constant low-energy contribution plus a linearly increasing component above approximately $\gamma$90 (Pfeiffer et al., 5 Sep 2025). BPT shows strong plasmon-energy dependence because its LUMO lies approximately $\gamma$91 above $\gamma$92, whereas ATP shows negligible change between $\gamma$93 and $\gamma$94, consistent with nonresonant interfacial scattering (Stefancu et al., 16 Jul 2025). By contrast, the Au oxidation study, limited to the near-IR intraband window, demonstrates strong CID but cannot determine whether its microscopic origin is predominantly roughness-like, charge-transfer-like, or mixed (Pfeiffer et al., 2024).

Second, CID can be large enough to become a major term in the effective damping rate. In nanostructures with large surface-to-volume ratio, it can be a dominant contribution, and its energy dependence implies stronger damping at shorter wavelengths once $\gamma$95 exceeds an interfacial threshold such as $\gamma$96 (Pfeiffer et al., 5 Sep 2025). In gap-waveguides, strong confinement enhances surface-related channels, making CID readily observable as a significant fraction of total loss (Stefancu et al., 16 Jul 2025). For oxidized planar Au, $\gamma$97 up to $\gamma$98 was sufficient to imply a substantial drop in plasmon quality factor (Pfeiffer et al., 2024).

Third, CID has direct implications for hot carriers and photoelectrochemistry. The decanethiol study states that the direct charge-transfer contribution is particularly relevant for hot-carrier generation and photocatalysis because it provides an interfacial pathway that can selectively enhance carrier injection into chemical or catalytic sites (Pfeiffer et al., 5 Sep 2025). The waveguide study similarly distinguishes CID that occurs within the SPR lifetime, approximately $\gamma$99–$$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$00, from three-step hot-carrier processes that occur after SPR dephasing (Stefancu et al., 16 Jul 2025). This suggests that CID is not merely a loss channel; under suitable energy-level alignment it is also a direct route for interfacial energy conversion.

Several limitations and controversies remain. In the decanethiol study, the decoupling of roughness and charge transfer relies on the spectral signature of a low-energy constant offset plus a linear rise above approximately $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$01; the authors note that roughness could have weak residual energy dependence and that charge transfer may start gradually rather than as a sharp threshold (Pfeiffer et al., 5 Sep 2025). In the waveguide study, Persson’s model captures trends but underestimates BPT’s resonant CID, and jointly treating coherent and incoherent electron transfer remains an open challenge (Stefancu et al., 16 Jul 2025). In the oxidation study, previous ellipsometry work had attributed reflectance changes solely to oxide optical properties; the new analysis argues instead that CID is the primary cause of broadening when one models a dielectric oxide overlayer and allows the Au collision frequency to vary (Pfeiffer et al., 2024).

Landau damping is also explicitly addressed in the decanethiol/Au(111) system. Because plane-wave excitation of a flat film produces very small fields normal to the surface inside the metal, with $$\Gamma_{\mathrm{tot}}=\Gamma_{\mathrm{rad}}+\Gamma_{\mathrm{bulk}}+\Gamma_{\mathrm{surf}}+\Gamma_{\mathrm{CID}},$$02, the measured energy-dependent increase was not attributed to Landau effects (Pfeiffer et al., 5 Sep 2025). This is important because CID is often discussed alongside bulk and surface nonradiative losses, and misassignments between these channels can affect physical interpretation.

The present literature therefore supports a broad but technically specific definition of CID: it is an interfacial augmentation of electronic damping that can arise from chemistry-induced roughness, nonresonant inelastic scattering, or resonant/direct charge transfer, and whose observable form depends on spectral window, field geometry, adsorbate electronic structure, and interfacial hybridization. A plausible implication is that CID should be treated not as a single phenomenological constant, but as an interface-specific function of photon energy, local field polarization, and surface chemistry.